Negotiating Search
Vertical search is not a new area of web apps but I learned something new about it recently. I attended a panel of CTO’s from the space yesterday at the Silicon Valley Web Builder event. Participating companies included Simply Hired, Spock, MEDgle and Riya. Each representative took a couple of minutes to conduct a product demonstration. Spock is still in a private beta, so it was neat to see it in action . My first impressions were very positive.
What’s interesting is that even though these are all search companies, none of the products are competitive. You might think this a very natural dynamic. After all, each vertical is necessarily focused on different slices of web content. But there’s more that makes search a different app as you move from vertical to vertical. It’s not simply a matter of different information. For me, the dividing line between these apps is rooted in how users interact with that information. The way the search process is organized to help users negotiate that interaction can vary widely from app to app.
Negotiating a search app has many factors to it. By negotiation, I mean interaction with flows of information. There are two major examples of how the applications mentioned earlier change the negotiation paradigm.
- Search as a Service – You seed a search on MedGLE by selecting a body part from a visual anatomy and the issue you are having with it. There are no keywords to initiate the process. This is a different input mechanism which is a filter rather than query. MedGLE is more of an expert system where you provide symptoms associated with where you feel them to allow it to run constraints against possible diagnosis. Those possible diagnosis are provided with some ranking. There’s a layer of logic there which extends beyond search. Meanwhile, Riya has developed Like.com to use visual search as enabling technology for a distinctive application around shopping. Like.com is interesting because it conforms to the way many people (especially women) shop, which is to window-shop or walk the isles of a store to see what’s eye-catching. Search is a second-order factor in the user experience. It’s more of an expressive shop-by-browsing experience. I would say the first-order elements of the user experience is browse and configure. Both companies demonstrate how the search interface is pushed a little lower in the stack (more important to other product components than to the user?).
- Search in the Big Picture – Simply Hired and Spock demonstrate search as a step in process-oriented activities. Simply Hired does a splendid job of organizing the needs of a job hunter into one application (sometimes one screen). By integrating third-party salary information, maps (figure out the commute) and LinkedIn (figure out who you know there), you can grow the information flow around a job very quickly. Or you could save a job posting to a clipboard-like space and add some notes. The user intent behind Spock is a lot less serious than job hunting (think vanity searches and entertainment) but broad similarities exist because search only kicks-off the experience and there a lot you might do after you find people. For example, if you retrieved results for Alex Rodriquez, a sometimes controversial athlete, you might be able to rate him or view news feeds that reveal his latest issue.
Now, it’s impossible to write about search apps without mentioning Google. Google has obviously habituated users to certain information negotiation paradigms. To contrast it with the above, let’s call it the “Search as Search” approach. It works splendidly for the most common and widespread user needs. But what the vertical search companies demonstrate is real diversity in user needs around information retrieval. For this reason, even Google’s new Universal Search is not as big a threat to these focused companies as you may think. That’s because certain needs are not rooted in search quality, which is Google’s forte. The challenge for vertical search companies, however, is that these needs arise from time to time which makes it hard to drive repeat usage. Monetization, therefore, relies on identifying high value categories (good news for Simply Hired and MedGLE) or those that straddle entertainment/fun-seeking motivations (Spock and Like.com) since those are more sticky habits to kick.
A possible idea for vertical providers is contextual integration. Imagine reading a story on a content site that mentions a person where Spock (or similar company) has been syndicated. When you read about A-Rod’s latest gaffe on ESPN.com or NY Times.com, you can deposit your rant or rating via Spock, which becomes the system-of-record for people info no matter where people are being discussed or covered.
Related Links
- Impact of Universal Search
- Chatting with the MEDgle.com Founders
- First Look at Spock
- Google May Employ Simply Hired
Filed under: Events, Search Technology, User Experiences, Web Apps
Email It
Digg This!
Slashdot It!
Tags: Vertical Search, Like, Riya, Simply Hired, Spock, MEDgle,

I'd be interested to see which sectors vertical search makes the most sense and whether any of the early companies are making significant headway attracting users. For instance, with specialized language and the need for trust, health would seem to be a natural, but I have yet to see Healthline.com break out of the pack. Are there other examples?